| ??? 11/03/00 18:28 Read: times |
#6195 - RE: Memory Test Algorithm |
Sorry Peter, but it is not that simple. This would take to much time.
The thing about the walking one's and Zero's is that you do not need to test the whole memory, meaning "all" locations. you just "test" some special - calculated - adresses and check in the second pass if you find the stuff where it should / or should not be stored. I have seen this algorithm checking 1MByte RAM incredible fast. it must work like this : maybe first clean whole memory, but not shure about that .... storing 0x01 @ 0x0001 storing 0x02 @ 0x0002 storing 0x04 @ 0x0004 storing 0x08 @ 0x0008 ... until .... end of memory .... storing 0x80 @ 0x8000 now checking 0x0001 == 0x01 checking 0x0003 != 0x01 checking 0x0005 != 0x01 checking 0x0009 != 0x01 see that "1" walking ? until checking 0x8001 != 0x01 now checking 0x0002 == 0x02 checking 0x0003 != 0x02 checking 0x0005 != 0x02 checking 0x0009 != 0x02 see that "1" walking too ? so just only a few checks now the whole stuff inversed with a walking "0" or maybe it was some mix like walking 1 on adressline and walking 0 at the dataline, and in the next pass inverted ... i am right, this will check any shurt's in the adresslines and datalines. but i have lost all documents about it, poor boy i am. maybe you see the point that this algorithm does not check the whole memory, but it checks the produced board for a very fast short circuit test. |
| Topic | Author | Date |
| Memory Test Algorithm | 01/01/70 00:00 | |
| RE: Memory Test Algorithm | 01/01/70 00:00 | |
| RE: Memory Test Algorithm | 01/01/70 00:00 | |
| RE: Memory Test Algorithm | 01/01/70 00:00 | |
| RE: Memory Test Algorithm | 01/01/70 00:00 | |
| RE: Memory Test Algorithm | 01/01/70 00:00 | |
RE: Memory Test Algorithm | 01/01/70 00:00 |



